Epidemiology of Pneumocystis carinii pneumonia in an era of effective prophylaxis: the relative contribution of non-adherence and drug failure

Objective: To determine the relative contribution of patient non-adherence, provider failure to prescribe prophylaxis, and drug failure to the continu ed occurrence of Pneumocystis carinii pneumonia (PCP), and to determine cor relates of non-adherence. Design: Retrospective case-control study. Methods: Patients with confirmed or presumptive PCP from May 1995 to Septem ber 1997 who had at least 6 months of prior HIV care (cases) were compared to controls matched for initial CD4 cell count and date of initial HIV care . Results: The incidence of PCP declined by 85% in the 28 months of the study . Of the 118 cases of PCP identified, 59 (50%) were in HIV care for > 6 mon ths prior to PCP diagnosis. In a multivariate logistic regression model, ri sk factors for PCP among patients in HIV care were patient non-adherence [o dds ratio (OR), 12.4; 95% confidence interval (CI), 6.4-23.5], use of proph ylaxis other than trimethoprimsulfamethoxazole (OR, 27.0; 95% CI, 13.8-52.9 ), and absence of antiretroviral use (OR, 7.5; 95% CI, 4.5-12.5). Provider non-adherence occurred in one out of 59 cases (2%), and five out of 106 con trols (5%). Of the patients who developed PCP on prophylaxis, 18 cases (30% ) appeared due to drug failure; there were no cases of apparent drug failur e among patients on trimethoprim-sulfamethoxazole. In multivariate analysis , non-adherence was more common among patients of non-white race, those wit h a history of injecting drug use, and those with active substance abuse or psychiatric illness. Conclusions: Patient non-adherence was the most common reason for the occur rence of PCP among patients in HIV care; provider non-adherence was uncommo n. Drug failure occurred only among patients on prophylaxis other than trim ethoprim-sulfamethoxazole. (C) 2000 Lippincott Williams & Wilkins.